Pith. sign in

REVIEW 6 major objections 7 minor 62 references

Enhancing Drug Discovery: Autoencoder-Based Latent Space Augmentation for Improved Molecular Solubility Prediction using LatMixSol

T0 review · 6 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Mixing molecule codes in latent space cuts solubility prediction error by up to 7.6%.

desk verdict A plausible but sloppily evaluated augmentation trick whose headline numbers don't match its own table; worth a serious referee only because the idea is cheap to test. read the letter →

arxiv 2506.00223 v1 pith:TYHLUMYL submitted 2025-05-30 q-bio.QM cs.LG

Mohammad Saleh Hasankhani This is my paper
classification q-bio.QMcs.LG
keywords LatMixSolmolecularsolubilitypredictiondataaugmentationspectralclusteringautoencoderlatentspaceUpinterpolationHuuskonendatasetgradientboosting
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes LatMixSol, a data-augmentation method that creates synthetic training samples by blending compressed descriptions of chemically similar compounds. It claims this reduces prediction error on a standard solubility benchmark for three of four gradient-boosted models, with RMSE drops of 3.2–7.6% and $R^2$ gains up to 1.5%, and reports the gains are statistically significant. The aim is to show that data scarcity in early drug discovery can be partially offset by generating plausible synthetic samples without needing extra experiments. If the method works, it offers a cheap way to improve solubility models in settings where labeled data are hard to obtain.

What carries the argument

The load-bearing object is the autoencoder's latent space with cluster-restricted MixUp. Molecules are encoded to $z\in\mathbb{R}^{64}$; spectral clustering with $k=10$ is done in the original descriptor space; and new samples are made by $z_{\mathrm{new}} = \lambda z_a + (1-\lambda) z_b$, with the same $\lambda$ blending the labels. Decoding returns a synthetic feature vector. The argument depends on the decoder being smooth enough that interpolation in latent space maps to chemically meaningful interpolation in feature space, and the paper explicitly notes that no chemical constraint is enforced.

What would settle it

Take a sample of the synthetic feature vectors generated by the method and find each one's nearest real molecule in the original descriptor space; then compare the interpolated solubility label with that neighbor's measured solubility. If the label errors approach the roughly $\pm 0.5$ log-unit noise of experimental solubility data, the interpolation-label assumption is not supported, and a stricter check would test whether any decoded vector corresponds to a chemically valid, synthesizable structure at all.

Watch

Extended reading notes

Core claim

LatMixSol's central claim is that cluster-guided interpolation in the latent space of an autoencoder produces chemically plausible solubility training data that improves downstream regression. The pipeline compresses 204 molecular descriptors to 64 variables, groups the training set into ten spectral clusters, and within each cluster blends pairs of latent vectors with a random weight $\lambda$, then decodes the blend and assigns it the interpolated solubility label $\lambda y_a + (1-\lambda) y_b$. On the Huuskonen benchmark (884 training, 413 test compounds), this 10x expansion lowers RMSE for CatBoost, LightGBM, and HistGradientBoosting while leaving XGBoost essentially unchanged; HistGradientBoosting improves most, by 7.6%. The paper also reports a maximum mean discrepancy of 0.0013 and low reconstruction error as evidence that the synthetic samples resemble the original data.

Load-bearing premise

The load-bearing assumption is that averaging two molecules' latent codes and averaging their solubility values produces a real compound with that averaged solubility, yet the paper states that no chemical constraint is enforced and offers no external check of the decoded structures.

Editorial extensions

If this is right

  • If the gains hold, augmenting descriptor space is a viable alternative to generating new molecular graphs for data-limited solubility modeling.
  • The improvement pattern suggests regularized gradient-boosted models benefit more from synthetic diversity; XGBoost's flat result is a warning that augmentation gains are model-dependent.
  • The reported reduction in cross-fold RMSE variance implies latent-space augmentation acts partly as a regularizer, stabilizing training on small datasets.
  • The low MMD and preserved pairwise distances, if accurate, mean the augmented set stays close to the original distribution rather than drifting into implausible regions.
  • SHAP analysis in the paper indicates the augmented models still lean on chemically meaningful descriptors such as lipophilicity, molecular weight, and complexity, preserving interpretability.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: the 'chemically valid' conclusion is stronger than its evidence, since interpolation is unconstrained by any chemical rule; a natural extension is to validate by checking decoded samples against real molecular structures or their nearest neighbors.
  • If interpolated labels are only approximately right, the method could be extended to pseudo-labeling or uncertainty-weighted blending to reduce noise from mismatched synthetic labels.
  • The same cluster-guided latent blending could be tested on other scarce ADME endpoints and larger curated solubility sets, though the transfer is not automatic.
  • A direct comparison against SMILES enumeration and graph perturbation on the same benchmark would clarify when latent-space augmentation is actually better; the paper does not run that comparison.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

6 major / 7 minor

Summary. The manuscript proposes LatMixSol, a data-augmentation framework for aqueous solubility prediction. Starting from 204 RDKit descriptors for the Huuskonen dataset, the method trains a two-layer autoencoder, applies spectral clustering (k=10) in the original feature space, performs MixUp-style interpolation between latent codes within each cluster, and decodes the interpolated vectors to produce synthetic samples with interpolated labels. Four gradient-boosted regressors (CatBoost, LightGBM, HistGradientBoosting, XGBoost) are compared with and without augmentation via five-fold cross-validation. The paper claims consistent RMSE reductions of 3.2–7.6% for three models with paired t-test p<0.01, chemical validity of the synthetic data based on MMD and reconstruction error, and interpretable SHAP analyses.

Significance. If the performance and validity claims were fully substantiated, the paper would offer a plausible, low-cost augmentation strategy for descriptor-based solubility prediction and would complement existing SMILES/graph augmentation methods. The proposed pipeline is clearly described, uses a standard benchmark, and reports several diagnostics (MMD, nearest-neighbor distances, reconstruction error) beyond a single accuracy number. These are real strengths. However, the central quantitative claims are currently unsupported: the headline improvement range is inconsistent with Table 2, no inferential statistics are provided, no ablation isolates cluster-guided interpolation, and the chemical-validity evidence is self-referential. The contribution is therefore not yet established at the level required for publication.

major comments (6)
  1. [§1, §4.2, Table 2] The abstract and Section 4.2 state RMSE reductions of 3.2–7.6% and R2 increases of 0.5–1.5%, but the reductions implied by Table 2 are CatBoost 1.89%, LightGBM 6.07%, HistGradientBoosting 7.57%, and XGBoost −0.08%; the 'Final Model Evaluation' paragraph itself says 1.9–7.6%, so the manuscript contains two mutually inconsistent versions of the headline result, and this discrepancy must be resolved before any significance claim can be assessed.
  2. [§2, §3.5, §4.2] The paired t-test with p<0.01 is claimed without any test statistic, degrees of freedom, per-fold RMSE values, or error bars; with K=5 folds the test would have only 4 degrees of freedom, and the mean CatBoost improvement of 0.0107 in RMSE cannot be judged significant without the per-fold variance, which is not reported. No code, data, or autoencoder hyperparameters (learning rate, epochs, batch size, activation, regularization) are provided, so the Table 2 numbers cannot be independently reproduced.
  3. [§3.4–3.5] The experimental protocol does not state whether the autoencoder, the spectral clustering, and the synthetic-sample generation are fitted separately inside each cross-validation training fold. If these components are trained on all 1,297 compounds before the five-fold splits, information from the validation folds can enter the augmented training data indirectly, which would invalidate the held-out comparison; the paper should specify and, if necessary, correct the fold-aware procedure.
  4. [§3.4, §4.2] The chemical-validity claim is not established by the reported evidence. Section 3.4 explicitly states that 'no explicit chemical constraint is enforced,' and the metrics in Table 1 (MMD, nearest-neighbor distance, pairwise distance, reconstruction MSE) are all computed on synthetic samples generated by the same autoencoder; these measure distributional self-consistency, not chemical validity. The Related Work section promises 'RDKit-based structural checks,' but no such checks, no SMILES-level validation, and no external property comparison are reported in Section 4.
  5. [§3.4] No ablation or baseline isolates the contribution of cluster-guided latent interpolation. The paper does not compare against random MixUp in the original feature space, interpolation without spectral clustering, or interpolation across randomly selected pairs, so the claim that cluster-restricted interpolation is responsible for the observed improvements is not supported by the experiments.
  6. [§3.4] The synthetic label y_new = λ y_a + (1−λ) y_b is assumed to be the correct solubility value of the decoded x_new, but no evidence is provided that solubility is linear along latent-space interpolation paths; the Lipschitz-continuity argument only bounds feature-space smoothness, not label correctness, so errors in this assumption would propagate directly into the augmented training targets.
minor comments (7)
  1. [References] References [3] and [50] are explicitly flagged in the bibliography as 'Assumed reference' or 'not found'; these are not citations and must be replaced with verifiable sources before publication.
  2. [Figures] Figure order is confusing: Figures 3 and 4 are discussed in Section 4.1 before Figure 2 appears, and Figure 1 is referenced in Section 3.3 without a caption description of the training details.
  3. [§3.5] Section 3.5 calls the cross-validation 'stratified K-fold' for a regression task, but no stratification on the continuous target is defined; clarify whether this is standard KFold with a fixed seed.
  4. [§3.3–3.4] The symbol L is used for the autoencoder loss in Section 3.3 and for the Lipschitz constant in Section 3.4; use distinct notation to avoid ambiguity.
  5. [§1] The paper claims an 'open-source implementation' but provides no repository URL or release information; either provide the link and hyperparameters or remove the claim.
  6. [§4.2, Table 1] The nearest-neighbor distance results in Table 1 are reported without comparison to the NN distance distribution of the original data, so the statement that 'balanced neighbor distances' indicate success is not supported.
  7. [§5.2] The computational-efficiency claim ('10× data expansion in under 30 minutes using consumer-grade GPUs') is stated without details of the hardware, software versions, or measurement procedure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central performance claim is tested on a held-out external benchmark and does not reduce to its inputs.

full rationale

The central claim—that LatMixSol augmentation improves RMSE/R² for gradient-boosted regressors—is evaluated on the held-out Huuskonen test split (Section 3.5, Table 2), so it is not definitionally tied to the training procedure. The augmentation pipeline (autoencoder, spectral clustering, latent MixUp) is an assumed method, not derived from the target result. The chemical-validity claim is weaker: Section 4.2 supports it with MMD, nearest-neighbor distance, and reconstruction MSE computed on the same autoencoder-generated samples, which are self-consistency checks rather than external structural validation; the paper itself admits in Section 3.4 that "no explicit chemical constraint is enforced." This is an evidentiary weakness and a correctness risk, but it is not a circular derivation: no fitted parameter is renamed as a prediction, no load-bearing self-citation chain is used, and no equation reduces to another by construction. The abstract/Table 2 inconsistency (3.2–7.6% vs 1.9–7.6%) and the unreported paired t-test details are reproducibility and supportability issues, not circularity.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the assumption that interpolating in the latent space of a plain autoencoder yields chemically plausible molecules with correctly interpolated solubility labels. That assumption is stated in Section 3.4 but not externally validated. The method also relies on several hand-chosen constants (k=10, beta=10, latent dimension 64) and on a Lipschitz continuity property of the decoder that is asserted without proof. The reference to prior work [3] is unverified, so the actual novelty margin is uncertain.

free parameters (6)
  • Number of spectral clusters k = 10
    Set heuristically; no sensitivity analysis is provided, and the paper itself lists it as a limitation.
  • Augmentation multiplier beta = 10
    Chosen to generate 10x synthetic samples per cluster; no justification or ablation.
  • Autoencoder latent dimension = 64
    Fixed architecture choice; no comparison of other latent sizes.
  • RBF kernel width sigma (spectral clustering) = Not reported
    Sigma is required to build the affinity matrix in Section 3.4, but its value or selection method is omitted.
  • High-correlation filter threshold = 0.95
    Ad hoc threshold for removing correlated descriptors; no sensitivity analysis.
  • Autoencoder training hyperparameters (epochs, learning rate, batch size) = Not reported
    The paper states Adam optimizer and MSE loss but omits training duration, learning rate, regularization, and batch size, all required for reproduction.
assumptions (3)
  • domain assumption The decoder is Lipschitz continuous
    Section 3.4 states this assumption to justify smooth latent interpolation; no bound or verification is given.
  • ad hoc to paper Latent interpolation produces chemically valid, label-correct synthetic samples
    Core assumption of the augmentation; the paper admits no explicit chemical constraint is enforced and does not validate with external structural checks.
  • domain assumption The Huuskonen train/test split is based on molecular topology and structural diversity
    Section 3.1 asserts this without evidence; the standard Huuskonen benchmark is commonly used with a fixed random split, and the autoencoder training split is not specified, creating potential leakage.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Drug Discovery: Autoencoder-Based Latent Space Augmentation for Improved Molecular Solubility Prediction using LatMixSol." pith.science (2026). https://pith.science/paper/TYHLUMYL

@misc{pith2026250600223,
  author       = {Pith},
  title        = {Pith review of: Enhancing Drug Discovery: Autoencoder-Based Latent Space Augmentation for Improved Molecular Solubility Prediction using LatMixSol},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TYHLUMYL}},
  note         = {Machine review of arXiv:2506.00223}
}
read the original abstract

Accurate prediction of molecular solubility is a cornerstone of early-stage drug discovery, yet conventional machine learning models face significant challenges due to limited labeled data and the high-dimensional nature of molecular descriptors. To address these issues, we propose LatMixSol, a novel latent space augmentation framework that combines autoencoder-based feature compression with guided interpolation to enrich training data. Our approach first encodes molecular descriptors into a low-dimensional latent space using a two-layer autoencoder. Spectral clustering is then applied to group chemically similar molecules, enabling targeted MixUp-style interpolation within clusters. Synthetic samples are generated by blending latent vectors of cluster members and decoding them back to the original feature space. Evaluated on the Huuskonen solubility benchmark, LatMixSol demonstrates consistent improvements across three of four gradient-boosted regressors (CatBoost, LightGBM, HistGradientBoosting), achieving RMSE reductions of 3.2-7.6% and R-squared increases of 0.5-1.5%. Notably, HistGradientBoosting shows the most significant enhancement with a 7.6% RMSE improvement. Our analysis confirms that cluster-guided latent space augmentation preserves chemical validity while expanding dataset diversity, offering a computationally efficient strategy to enhance predictive models in resource-constrained drug discovery pipelines.

Figures

Figures reproduced from arXiv: 2506.00223 by the authors.

Figure 1
Figure 1. Autoencoder training loss L vs. epoch. Cluster-Specific MixUp in Latent Space For each cluster Ci (assigned in the original space), synthetic samples are generated as follows: 1. Randomly select two distinct samples xa, xb ∈ Ci . 2. Encode the samples into latent space: za = encoder(xa), zb = encoder(xb). 3. Perform linear interpolation with λ ∼ U[0, 1]: znew = λza + (1 − λ)zb. 4. Decode the new latent vector to obt… view at source ↗
Figure 3
Figure 3. Per-iteration RMSE using LatMixSol-based data augmentation. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Per-iteration RMSE without LatMixSol augmentation. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figures from the paper (3 more)
Figure 2
Figure 2. Figure 2: Block diagram illustrating the main steps of the methodology. [PITH_FULL_IMAGE:figures/full_fig_p009_2.png]
Figure 5
Figure 5. Figure 5: SHAP summary plots comparing feature importance for LightGBM and CatBoost with LatMixSol augmenta [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: SHAP dependency plots for LightGBM and CatBoost. [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

62 extracted references · 61 canonical work pages

  1. [1]

    Prediction of organic compound aqueous solubility using machine learning: A comparison study of descriptor-based and fingerprints-based models

    Soroush Ahmadi, Mohammad Amin Ghanavati, and Sohrab Rohani. Prediction of organic compound aqueous solubility using machine learning: A comparison study of descriptor-based and fingerprints-based models. Journal of Cheminformatics, 15:103, 2023

    work page 2023

  2. [2]

    Bai et al

    Y . Bai et al. Machine learning-enabled drug-induced toxicity prediction.Advanced Science, 12:e2407920, 2025

    work page 2025

  3. [3]

    Bhattacharjee et al

    S. Bhattacharjee et al. Spectral clustering and latent space interpolation for molecular data augmentation, 2023. Assumed reference for Bhattacharjee et al., as specific title not provided

    work page 2023

  4. [4]

    A hybrid approach to aqueous solubility prediction using cosmo-rs and machine learning

    Esben Jannik Bjerrum et al. A hybrid approach to aqueous solubility prediction using cosmo-rs and machine learning. Journal of Molecular Liquids, 375:121305, 2023

    work page 2023

  5. [5]

    Machine learning with physicochemical relationships: Solubility prediction in organic solvents and water

    Samuel Boobier et al. Machine learning with physicochemical relationships: Solubility prediction in organic solvents and water. Nature Communications, 11:5753, 2020

    work page 2020

  6. [6]

    Chen et al

    T. Chen et al. Data-efficient gan training beyond (just) augmentations: A lottery ticket perspective. Advances in Neural Information Processing Systems, 34:20941–20955, 2021

    work page 2021

  7. [7]

    Real external predictivity of qsar models: How to evaluate it? comparison of different validation criteria and proposal of using the concordance correlation coefficient

    Nicola Chirico and Paola Gramatica. Real external predictivity of qsar models: How to evaluate it? comparison of different validation criteria and proposal of using the concordance correlation coefficient. Journal of Chemical Information and Modeling, 51(9):2320–2335, 2011

    work page 2011

  8. [8]

    Cui et al

    Q. Cui et al. Evaluation of machine learning models for aqueous solubility prediction in drug discovery. bioRxiv, 2024

    work page 2024

Show all 62 references
  1. [9]

    Improved prediction of aqueous solubility of novel compounds by going deeper with deep learning

    Qiang Cui, Sijie Lu, Bing Ni, Xiang Zeng, Yajun Tan, Yuanchun Chen, and Hongmei Zhao. Improved prediction of aqueous solubility of novel compounds by going deeper with deep learning. Frontiers in Oncology, 10:121, 2020

    2020

  2. [10]

    John S. Delaney. Esol: Estimating aqueous solubility directly from molecular structure. Journal of Chemical Information and Modeling, 44(3):1000–1005, 2004

    2004

  3. [11]

    Ding et al

    F. Ding et al. Advancing aqueous solubility prediction: A machine learning approach for organic compounds using a curated dataset. ChemRxiv, 2024

    2024

  4. [12]

    Duchowicz and Eduardo A

    Pablo R. Duchowicz and Eduardo A. Castro. Qspr studies on aqueous solubilities of drug-like compounds. International Journal of Molecular Sciences, 10(6):2558–2577, 2009

    2009

  5. [13]

    Convolutional networks on graphs for learning molecular fingerprints

    David Duvenaud et al. Convolutional networks on graphs for learning molecular fingerprints. Advances in Neural Information Processing Systems, 28:2224–2232, 2015

    2015

  6. [14]

    Worth, Mark T

    Lennart Eriksson, Joanna Jaworska, Andrew P. Worth, Mark T. D. Cronin, Robert M. McDowell, and Paola Gramatica. Methods for reliability and uncertainty assessment and for applicability evaluations of classification and regression-based qsars. Environmental Health Perspectives,...

    2003

  7. [15]

    J. H. Fagerberg et al. Prediction of solubility in biorelevant media using pls methodology and dragonx descriptors. European Journal of Pharmaceutical Sciences, 78:232–240, 2015. 15 A PREPRINT - J ULY 4, 2025

    2015

  8. [16]

    P. G. Francoeur and D. R. Koes. Soltrannet–a machine learning tool for fast aqueous solubility prediction. Journal of Chemical Information and Modeling, 61(6):2530–2536, 2021

    2021

  9. [17]

    S. Gan, D. A. Cosgrove, E. J. Gardiner, and V . J. Gillet. Investigation of the use of spectral clustering for the analysis of molecular data. Journal of Chemical Information and Modeling, 54(12):3302–3310, 2014

    2014

  10. [18]

    Saldanha, Vijayakumar Murugesan, and Wei Wang

    Peiyuan Gao, Alexander Andersen, Jose Sepulveda, Gihan Panapitiya, Aaron Hollas, Edward G. Saldanha, Vijayakumar Murugesan, and Wei Wang. Somas: A platform for data-driven material discovery in redox flow battery development. Scientific Data, 9:736, 2022

    2022

  11. [19]

    A machine learning approach for the prediction of aqueous solubility of pharmaceuticals: A comparative model and dataset analysis

    Mohammad Amin Ghanavati, Soroush Ahmadi, and Sohrab Rohani. A machine learning approach for the prediction of aqueous solubility of pharmaceuticals: A comparative model and dataset analysis. Digital Discovery, 3:1845–1858, 2024

    2024

  12. [20]

    Wei, David Duvenaud, José Miguel Hernández-Lobato, Benjamin Sánchez- Lengeling, Dennis Sheberla, Jorge Aguilera-Iparraguirre, Timothy D

    Rafael Gómez-Bombarelli, Jennifer N. Wei, David Duvenaud, José Miguel Hernández-Lobato, Benjamin Sánchez- Lengeling, Dennis Sheberla, Jorge Aguilera-Iparraguirre, Timothy D. Hirzel, Ryan P. Adams, and Alán Aspuru- Guzik. Automatic chemical design using a data-driven continuous...

    2018

  13. [21]

    Hadipour, C

    H. Hadipour, C. Liu, R. Davis, S. T. Cardona, and P. Hu. Deep clustering of small molecules at large-scale via variational autoencoder embedding and k-means. BMC Bioinformatics, 23:123, 2022

    2022

  14. [22]

    Strategies for pre-training graph neural networks

    Weihua Hu et al. Strategies for pre-training graph neural networks. International Conference on Learning Representations, 2020

    2020

  15. [23]

    Estimation of aqueous solubility for a diverse set of organic compounds based on molecular topology

    Jarmo Huuskonen. Estimation of aqueous solubility for a diverse set of organic compounds based on molecular topology. Journal of Chemical Information and Computer Sciences, 40(3):773–777, 2000

    2000

  16. [24]

    Mix-key: Graph mixup with key structures for molecular property prediction

    Tianyi Jiang, Zeyu Wang, et al. Mix-key: Graph mixup with key structures for molecular property prediction. Briefings in Bioinformatics, 25(3):bbae165, 2024

    2024

  17. [25]

    Keith et al

    John A. Keith et al. Latent variable machine learning framework for catalysis: General models, transfer learning, and interpretability. JACS Au, 1(10):1686–1699, 2021

    2021

  18. [26]

    Comprehensive ensemble in qsar prediction for drug discovery

    Joseph Kwon, Sangsoo Lee, and Sun Kim. Comprehensive ensemble in qsar prediction for drug discovery. BMC Bioinformatics, 20:554, 2019

    2019

  19. [27]

    Conditional molecular design with deep generative models

    Jaechang Lim, Seokho Ryu, Jin Woo Kim, and Woo Youn Kim. Conditional molecular design with deep generative models. Journal of Chemical Information and Modeling, 58(1):24–35, 2018

    2018

  20. [28]

    C. A. Lipinski et al. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews, 46(1-3):3–26, 2001

    2001

  21. [29]

    Predictive models of aqueous solubility of organic compounds built on a large dataset of high integrity

    Antonio Llinas et al. Predictive models of aqueous solubility of organic compounds built on a large dataset of high integrity. Molecular Pharmaceutics, 19(7):2239–2254, 2022

    2022

  22. [30]

    Will we ever be able to accurately predict solubility? Scientific Data, 11:496, 2024

    Pol Llompart, Christian Minoletti, Sapark Baybekov, Dragos Horvath, Gilles Marcou, and Alexandre Varnek. Will we ever be able to accurately predict solubility? Scientific Data, 11:496, 2024

    2024

  23. [31]

    Lovatt et al

    Christopher J. Lovatt et al. Revisiting the application of machine learning approaches in predicting aqueous solubility. ACS Omega, 9(7):8056–8066, 2024

    2024

  24. [32]

    Mario Lovri´c, Kristina Pavlovi´c, Petar Žuvela, Adrian Spataru, Bono Luˇci´c, Roman Kern, and Ming Wah Wong. Machine learning in prediction of intrinsic aqueous solubility of drug-like compounds: Generalization, complexity, or predictive ability? Journal of Chemometrics, 35(7...

    2021

  25. [33]

    Lu et al

    J. Lu et al. Creation and interpretation of machine learning models for aqueous solubility prediction. Exploration of Drug and Medicine, 1:252–265, 2023

    2023

  26. [34]

    Luo et al

    Z. Luo et al. Prediction of the aqueous solubility of compounds based on light gradient boosting machines with molecular fingerprints and the cuckoo search algorithm. ACS Omega, 7(17):14462–14471, 2022

    2022

  27. [35]

    Deep architectures and deep learning in cheminformatics: The prediction of aqueous solubility for drug-like molecules

    Alessandro Lusci, Gianluca Pollastri, and Pierre Baldi. Deep architectures and deep learning in cheminformatics: The prediction of aqueous solubility for drug-like molecules. Journal of Chemical Information and Modeling, 53(7):1563–1575, 2013

    2013

  28. [36]

    Auglichem: Data augmentation for molecular property prediction

    Rishikesh Magar, Yuxuan Wang, Yuzhe Wang, and Chenliang Ding. Auglichem: Data augmentation for molecular property prediction. arXiv preprint, page arXiv:2108.03356, 2021

    2021 arXiv

  29. [37]

    McMeekin et al

    H. McMeekin et al. Integrating machine learning and shap analysis to advance the rational design of benzoth- iadiazole derivatives with tailored photophysical properties. Journal of Physical Chemistry A, 129:1234–1245, 2025. 16 A PREPRINT - J ULY 4, 2025

    2025

  30. [38]

    D. S. Palmer and J. B. O. Mitchell. Is experimental data quality the limiting factor in predicting the aqueous solubility of druglike molecules? Molecular Pharmaceutics, 11(8):2962–2972, 2014

    2014

  31. [39]

    Pan et al

    L. Pan et al. Computational prediction of drug solubility in water-based systems: Qualitative and quantitative approaches used in the current drug discovery. AAPS PharmSciTech, 22:118, 2021

    2021

  32. [40]

    Evaluation of deep learning architectures for aqueous solubility prediction

    Gihan Panapitiya, Madeline Girard, Aaron Hollas, Jose Sepulveda, Vijayakumar Murugesan, Wei Wang, and Edward Saldanha. Evaluation of deep learning architectures for aqueous solubility prediction. ACS Omega, 7(28):24596–24608, 2022

    2022

  33. [41]

    Marcio Cerbella Ramos and Andrew D. White. Predicting small molecules solubility on endpoint devices using deep ensemble neural networks. Digital Discovery, 3:1094–1106, 2024

    2024

  34. [42]

    Reiter et al

    S. Reiter et al. Interpolatory model reduction of dynamical systems with root mean squared error. arXiv preprint arXiv:2403.08894, 2024

    2024 arXiv

  35. [43]

    Explainable machine learning for property predictions in compound optimization

    Raquel Rodríguez-Pérez and Jürgen Bajorath. Explainable machine learning for property predictions in compound optimization. Journal of Medicinal Chemistry, 64(22):16439–16451, 2021

    2021

  36. [44]

    Accurate, reliable and interpretable solubility prediction of druglike molecules with attention pooling and bayesian learning

    Seongok Ryu and Sumin Lee. Accurate, reliable and interpretable solubility prediction of druglike molecules with attention pooling and bayesian learning. arXiv preprint, page arXiv:2210.07145, 2022

    2022 arXiv

  37. [45]

    Molecular transformer: A model for uncertainty-calibrated chemical property prediction

    Philippe Schwaller, Teodoro Gaudin, David Lanyi, Costas Bekas, and Teodoro Laino. Molecular transformer: A model for uncertainty-calibrated chemical property prediction. ACS Central Science, 5(9):1572–1583, 2019

    2019

  38. [46]

    M. C. Sorkun et al. Boosting the predictive performance with aqueous solubility dataset curation. Scientific Data, 9:71, 2022

    2022

  39. [47]

    Aqsoldb, a curated reference set of aqueous solubility and 2d descriptors for a diverse set of compounds

    Murat Cihan Sorkun, Ashish Khetan, and Serkan Er. Aqsoldb, a curated reference set of aqueous solubility and 2d descriptors for a diverse set of compounds. Scientific Data, 6:151, 2019

    2019

  40. [48]

    Christopher Culberson, Robert P

    Vladimir Svetnik, Andy Liaw, Christopher Tong, J. Christopher Culberson, Robert P. Sheridan, and Bradley P. Feuston. Random forest: A classification and regression tool for compound classification and qsar modeling. Journal of Chemical Information and Computer Sciences, 43(6):...

    1947

  41. [49]

    Boosting the predictive performance with aqueous solubility dataset curation

    Fei Tang et al. Boosting the predictive performance with aqueous solubility dataset curation. Scientific Data, 9:71, 2022

    2022

  42. [50]

    Tevosyan et al

    A. Tevosyan et al. Rule-based data augmentation for chemical property prediction, 2023. Assumed reference for Tevosyan et al., as specific title not found

    2023

  43. [51]

    Deep learning methods in protein structure prediction

    Mauro Torrisi, Gianluca Pollastri, and Quynh Le. Deep learning methods in protein structure prediction. Compu- tational and Structural Biotechnology Journal, 18:1301–1310, 2020

    2020

  44. [52]

    Uzundurukan et al

    A. Uzundurukan et al. Meta-analysis and review of in silico methods in drug discovery – part 1: technological evolution and trends from big data to chemical space. Pharmacogenomics Journal, 25:8, 2025

    2025

  45. [53]

    Wang et al

    Y . Wang et al. Graphst: Spatially informed clustering, integration, and deconvolution of spatial transcriptomics with graph self-supervised contrastive learning. Nature Communications, 14:1048, 2023

    2023

  46. [54]

    Unsupervised learning methods for molecular simulation data

    Oliver Wieder et al. Unsupervised learning methods for molecular simulation data. Chemical Reviews , 121(18):12052–12086, 2021

    2021

  47. [55]

    Efficient clustering of large molecular libraries

    Chunxiao Yan et al. Efficient clustering of large molecular libraries. Journal of Chemical Information and Modeling, 60(12):6093–6101, 2020

    2020

  48. [56]

    Integrated molecular modeling and machine learning for drug design

    Kevin Yang et al. Integrated molecular modeling and machine learning for drug design. Journal of Chemical Theory and Computation, 17(6):3521–3533, 2021

    2021

  49. [57]

    Yang et al

    Y . Yang et al. Multi-channel gcn ensembled machine learning model for molecular aqueous solubility prediction on a clean dataset. Molecular Diversity, 27:2135–2145, 2023

    2023

  50. [58]

    Graph contrastive learning with augmentations

    Jiaxuan You et al. Graph contrastive learning with augmentations. Advances in Neural Information Processing Systems, 33:5812–5823, 2020

    2020

  51. [59]

    Zhang and R

    B. Zhang and R. Sennrich. Root mean square layer normalization. Advances in Neural Information Processing Systems, 32:12360–12371, 2019

    2019

  52. [60]

    Graph contrastive learning with augmentations

    Han Zhang et al. Graph contrastive learning with augmentations. International Conference on Learning Representations, 2021

    2021

  53. [61]

    Zhang et al

    W. Zhang et al. Predicting drug solubility using different machine learning methods - linear regression model with extracted chemical features vs graph convolutional neural network. arXiv, 2024

    2024

  54. [62]

    Zhang et al

    Z. Zhang et al. Cancer subtype identification by multi-omics clustering based on interpretable feature and latent subspace learning. Computers in Biology and Medicine, 152:106374, 2023. 17

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.